Add channel layout support to the AC-3 encoder.
[FFMpeg-mirror/lagarith.git] / libavcodec / libamr.c
blobe487a737f1debd2418c763ada62ba426f3517a6a
1 /*
2 * AMR Audio decoder stub
3 * Copyright (c) 2003 the ffmpeg project
5 * This file is part of FFmpeg.
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 /** @file
23 * Adaptive Multi-Rate (AMR) Audio decoder stub.
25 * This code implements both an AMR-NarrowBand (AMR-NB) and an AMR-WideBand
26 * (AMR-WB) audio encoder/decoder through external reference code from
27 * http://www.3gpp.org/. The license of the code from 3gpp is unclear so you
28 * have to download the code separately. Two versions exists: One fixed-point
29 * and one floating-point. For some reason the float encoder is significantly
30 * faster at least on a P4 1.5GHz (0.9s instead of 9.9s on a 30s audio clip
31 * at MR102). Both float and fixed point are supported for AMR-NB, but only
32 * float for AMR-WB.
34 * \section AMR-NB
36 * \subsection Float
37 * The float version (default) can be downloaded from:
38 * http://www.3gpp.org/ftp/Specs/archive/26_series/26.104/26104-610.zip
40 * \subsection Fixed-point
41 * The fixed-point (TS26.073) can be downloaded from:
42 * http://www.3gpp.org/ftp/Specs/archive/26_series/26.073/26073-600.zip
44 * \subsection Specification
45 * The specification for AMR-NB can be found in TS 26.071
46 * (http://www.3gpp.org/ftp/Specs/html-info/26071.htm) and some other
47 * info at http://www.3gpp.org/ftp/Specs/html-info/26-series.htm.
49 * \section AMR-WB
51 * \subsection Float
52 * The reference code can be downloaded from:
53 * http://www.3gpp.org/ftp/Specs/archive/26_series/26.204/26204-600.zip
55 * \subsection Fixed-point
56 * If someone wants to use the fixed point version it can be downloaded from:
57 * http://www.3gpp.org/ftp/Specs/archive/26_series/26.173/26173-571.zip.
59 * \subsection Specification
60 * The specification for AMR-WB can be found in TS 26.171
61 * (http://www.3gpp.org/ftp/Specs/html-info/26171.htm) and some other
62 * info at http://www.3gpp.org/ftp/Specs/html-info/26-series.htm.
66 #include "avcodec.h"
68 #if CONFIG_LIBAMR_NB_FIXED
70 #define MMS_IO
72 #include "amr/sp_dec.h"
73 #include "amr/d_homing.h"
74 #include "amr/typedef.h"
75 #include "amr/sp_enc.h"
76 #include "amr/sid_sync.h"
77 #include "amr/e_homing.h"
79 #else
80 #include <amrnb/interf_dec.h>
81 #include <amrnb/interf_enc.h>
82 #endif
84 static const char nb_bitrate_unsupported[] =
85 "bitrate not supported: use one of 4.75k, 5.15k, 5.9k, 6.7k, 7.4k, 7.95k, 10.2k or 12.2k\n";
86 static const char wb_bitrate_unsupported[] =
87 "bitrate not supported: use one of 6.6k, 8.85k, 12.65k, 14.25k, 15.85k, 18.25k, 19.85k, 23.05k, or 23.85k\n";
89 /* Common code for fixed and float version*/
90 typedef struct AMR_bitrates
92 int rate;
93 enum Mode mode;
94 } AMR_bitrates;
96 /* Match desired bitrate */
97 static int getBitrateMode(int bitrate)
99 /* make the correspondance between bitrate and mode */
100 AMR_bitrates rates[]={ {4750,MR475},
101 {5150,MR515},
102 {5900,MR59},
103 {6700,MR67},
104 {7400,MR74},
105 {7950,MR795},
106 {10200,MR102},
107 {12200,MR122},
109 int i;
111 for(i=0;i<8;i++)
113 if(rates[i].rate==bitrate)
115 return rates[i].mode;
118 /* no bitrate matching, return an error */
119 return -1;
122 static void amr_decode_fix_avctx(AVCodecContext * avctx)
124 const int is_amr_wb = 1 + (avctx->codec_id == CODEC_ID_AMR_WB);
126 if(avctx->sample_rate == 0)
128 avctx->sample_rate = 8000 * is_amr_wb;
131 if(avctx->channels == 0)
133 avctx->channels = 1;
136 avctx->frame_size = 160 * is_amr_wb;
137 avctx->sample_fmt = SAMPLE_FMT_S16;
140 #if CONFIG_LIBAMR_NB_FIXED
141 /* fixed point version*/
142 /* frame size in serial bitstream file (frame type + serial stream + flags) */
143 #define SERIAL_FRAMESIZE (1+MAX_SERIAL_SIZE+5)
145 typedef struct AMRContext {
146 int frameCount;
147 Speech_Decode_FrameState *speech_decoder_state;
148 enum RXFrameType rx_type;
149 enum Mode mode;
150 Word16 reset_flag;
151 Word16 reset_flag_old;
153 int enc_bitrate;
154 Speech_Encode_FrameState *enstate;
155 sid_syncState *sidstate;
156 enum TXFrameType tx_frametype;
157 } AMRContext;
159 static av_cold int amr_nb_decode_init(AVCodecContext * avctx)
161 AMRContext *s = avctx->priv_data;
163 s->frameCount=0;
164 s->speech_decoder_state=NULL;
165 s->rx_type = (enum RXFrameType)0;
166 s->mode= (enum Mode)0;
167 s->reset_flag=0;
168 s->reset_flag_old=1;
170 if(Speech_Decode_Frame_init(&s->speech_decoder_state, "Decoder"))
172 av_log(avctx, AV_LOG_ERROR, "Speech_Decode_Frame_init error\n");
173 return -1;
176 amr_decode_fix_avctx(avctx);
178 if(avctx->channels > 1)
180 av_log(avctx, AV_LOG_ERROR, "amr_nb: multichannel decoding not supported\n");
181 return -1;
184 return 0;
187 static av_cold int amr_nb_encode_init(AVCodecContext * avctx)
189 AMRContext *s = avctx->priv_data;
191 s->frameCount=0;
192 s->speech_decoder_state=NULL;
193 s->rx_type = (enum RXFrameType)0;
194 s->mode= (enum Mode)0;
195 s->reset_flag=0;
196 s->reset_flag_old=1;
198 if(avctx->sample_rate!=8000)
200 av_log(avctx, AV_LOG_ERROR, "Only 8000Hz sample rate supported\n");
201 return -1;
204 if(avctx->channels!=1)
206 av_log(avctx, AV_LOG_ERROR, "Only mono supported\n");
207 return -1;
210 avctx->frame_size=160;
211 avctx->coded_frame= avcodec_alloc_frame();
213 if(Speech_Encode_Frame_init(&s->enstate, 0, "encoder") || sid_sync_init (&s->sidstate))
215 av_log(avctx, AV_LOG_ERROR, "Speech_Encode_Frame_init error\n");
216 return -1;
219 if((s->enc_bitrate=getBitrateMode(avctx->bit_rate))<0)
221 av_log(avctx, AV_LOG_ERROR, nb_bitrate_unsupported);
222 return -1;
225 return 0;
228 static av_cold int amr_nb_encode_close(AVCodecContext * avctx)
230 AMRContext *s = avctx->priv_data;
232 Speech_Encode_Frame_exit(&s->enstate);
233 sid_sync_exit (&s->sidstate);
234 av_freep(&avctx->coded_frame);
235 return 0;
238 static av_cold int amr_nb_decode_close(AVCodecContext * avctx)
240 AMRContext *s = avctx->priv_data;
242 Speech_Decode_Frame_exit(&s->speech_decoder_state);
243 return 0;
246 static int amr_nb_decode_frame(AVCodecContext * avctx,
247 void *data, int *data_size,
248 AVPacket *avpkt)
250 const uint8_t *buf = avpkt->data;
251 int buf_size = avpkt->size;
252 AMRContext *s = avctx->priv_data;
253 const uint8_t*amrData=buf;
254 int offset=0;
255 UWord8 toc, q, ft;
256 Word16 serial[SERIAL_FRAMESIZE]; /* coded bits */
257 Word16 *synth;
258 UWord8 *packed_bits;
259 static Word16 packed_size[16] = {12, 13, 15, 17, 19, 20, 26, 31, 5, 0, 0, 0, 0, 0, 0, 0};
260 int i;
262 //printf("amr_decode_frame data_size=%i buf=0x%X buf_size=%d frameCount=%d!!\n",*data_size,buf,buf_size,s->frameCount);
264 synth=data;
266 toc=amrData[offset];
267 /* read rest of the frame based on ToC byte */
268 q = (toc >> 2) & 0x01;
269 ft = (toc >> 3) & 0x0F;
271 //printf("offset=%d, packet_size=%d amrData= 0x%X %X %X %X\n",offset,packed_size[ft],amrData[offset],amrData[offset+1],amrData[offset+2],amrData[offset+3]);
273 offset++;
275 packed_bits=amrData+offset;
277 offset+=packed_size[ft];
279 //Unsort and unpack bits
280 s->rx_type = UnpackBits(q, ft, packed_bits, &s->mode, &serial[1]);
282 //We have a new frame
283 s->frameCount++;
285 if (s->rx_type == RX_NO_DATA)
287 s->mode = s->speech_decoder_state->prev_mode;
289 else {
290 s->speech_decoder_state->prev_mode = s->mode;
293 /* if homed: check if this frame is another homing frame */
294 if (s->reset_flag_old == 1)
296 /* only check until end of first subframe */
297 s->reset_flag = decoder_homing_frame_test_first(&serial[1], s->mode);
299 /* produce encoder homing frame if homed & input=decoder homing frame */
300 if ((s->reset_flag != 0) && (s->reset_flag_old != 0))
302 for (i = 0; i < L_FRAME; i++)
304 synth[i] = EHF_MASK;
307 else
309 /* decode frame */
310 Speech_Decode_Frame(s->speech_decoder_state, s->mode, &serial[1], s->rx_type, synth);
313 //Each AMR-frame results in 160 16-bit samples
314 *data_size=160*2;
316 /* if not homed: check whether current frame is a homing frame */
317 if (s->reset_flag_old == 0)
319 /* check whole frame */
320 s->reset_flag = decoder_homing_frame_test(&serial[1], s->mode);
322 /* reset decoder if current frame is a homing frame */
323 if (s->reset_flag != 0)
325 Speech_Decode_Frame_reset(s->speech_decoder_state);
327 s->reset_flag_old = s->reset_flag;
329 return offset;
333 static int amr_nb_encode_frame(AVCodecContext *avctx,
334 unsigned char *frame/*out*/, int buf_size, void *data/*in*/)
336 short serial_data[250] = {0};
337 AMRContext *s = avctx->priv_data;
338 int written;
340 s->reset_flag = encoder_homing_frame_test(data);
342 Speech_Encode_Frame(s->enstate, s->enc_bitrate, data, &serial_data[1], &s->mode);
344 /* add frame type and mode */
345 sid_sync (s->sidstate, s->mode, &s->tx_frametype);
347 written = PackBits(s->mode, s->enc_bitrate, s->tx_frametype, &serial_data[1], frame);
349 if (s->reset_flag != 0)
351 Speech_Encode_Frame_reset(s->enstate);
352 sid_sync_reset(s->sidstate);
354 return written;
358 #elif CONFIG_LIBAMR_NB /* Float point version*/
360 typedef struct AMRContext {
361 int frameCount;
362 void * decState;
363 int *enstate;
364 int enc_bitrate;
365 } AMRContext;
367 static av_cold int amr_nb_decode_init(AVCodecContext * avctx)
369 AMRContext *s = avctx->priv_data;
371 s->frameCount=0;
372 s->decState=Decoder_Interface_init();
373 if(!s->decState)
375 av_log(avctx, AV_LOG_ERROR, "Decoder_Interface_init error\r\n");
376 return -1;
379 amr_decode_fix_avctx(avctx);
381 if(avctx->channels > 1)
383 av_log(avctx, AV_LOG_ERROR, "amr_nb: multichannel decoding not supported\n");
384 return -1;
387 return 0;
390 static av_cold int amr_nb_encode_init(AVCodecContext * avctx)
392 AMRContext *s = avctx->priv_data;
394 s->frameCount=0;
396 if(avctx->sample_rate!=8000)
398 av_log(avctx, AV_LOG_ERROR, "Only 8000Hz sample rate supported\n");
399 return -1;
402 if(avctx->channels!=1)
404 av_log(avctx, AV_LOG_ERROR, "Only mono supported\n");
405 return -1;
408 avctx->frame_size=160;
409 avctx->coded_frame= avcodec_alloc_frame();
411 s->enstate=Encoder_Interface_init(0);
412 if(!s->enstate)
414 av_log(avctx, AV_LOG_ERROR, "Encoder_Interface_init error\n");
415 return -1;
418 if((s->enc_bitrate=getBitrateMode(avctx->bit_rate))<0)
420 av_log(avctx, AV_LOG_ERROR, nb_bitrate_unsupported);
421 return -1;
424 return 0;
427 static av_cold int amr_nb_decode_close(AVCodecContext * avctx)
429 AMRContext *s = avctx->priv_data;
431 Decoder_Interface_exit(s->decState);
432 return 0;
435 static av_cold int amr_nb_encode_close(AVCodecContext * avctx)
437 AMRContext *s = avctx->priv_data;
439 Encoder_Interface_exit(s->enstate);
440 av_freep(&avctx->coded_frame);
441 return 0;
444 static int amr_nb_decode_frame(AVCodecContext * avctx,
445 void *data, int *data_size,
446 AVPacket *avpkt)
448 const uint8_t *buf = avpkt->data;
449 int buf_size = avpkt->size;
450 AMRContext *s = avctx->priv_data;
451 const uint8_t*amrData=buf;
452 static const uint8_t block_size[16]={ 12, 13, 15, 17, 19, 20, 26, 31, 5, 0, 0, 0, 0, 0, 0, 0 };
453 enum Mode dec_mode;
454 int packet_size;
456 /* av_log(NULL,AV_LOG_DEBUG,"amr_decode_frame buf=%p buf_size=%d frameCount=%d!!\n",buf,buf_size,s->frameCount); */
458 dec_mode = (buf[0] >> 3) & 0x000F;
459 packet_size = block_size[dec_mode]+1;
461 if(packet_size > buf_size) {
462 av_log(avctx, AV_LOG_ERROR, "amr frame too short (%u, should be %u)\n", buf_size, packet_size);
463 return -1;
466 s->frameCount++;
467 /* av_log(NULL,AV_LOG_DEBUG,"packet_size=%d amrData= 0x%X %X %X %X\n",packet_size,amrData[0],amrData[1],amrData[2],amrData[3]); */
468 /* call decoder */
469 Decoder_Interface_Decode(s->decState, amrData, data, 0);
470 *data_size=160*2;
472 return packet_size;
475 static int amr_nb_encode_frame(AVCodecContext *avctx,
476 unsigned char *frame/*out*/, int buf_size, void *data/*in*/)
478 AMRContext *s = avctx->priv_data;
479 int written;
481 if((s->enc_bitrate=getBitrateMode(avctx->bit_rate))<0)
483 av_log(avctx, AV_LOG_ERROR, nb_bitrate_unsupported);
484 return -1;
487 written = Encoder_Interface_Encode(s->enstate,
488 s->enc_bitrate,
489 data,
490 frame,
492 /* av_log(NULL,AV_LOG_DEBUG,"amr_nb_encode_frame encoded %u bytes, bitrate %u, first byte was %#02x\n",written, s->enc_bitrate, frame[0] ); */
494 return written;
497 #endif
499 #if CONFIG_LIBAMR_NB || CONFIG_LIBAMR_NB_FIXED
501 AVCodec libamr_nb_decoder =
503 "libamr_nb",
504 CODEC_TYPE_AUDIO,
505 CODEC_ID_AMR_NB,
506 sizeof(AMRContext),
507 amr_nb_decode_init,
508 NULL,
509 amr_nb_decode_close,
510 amr_nb_decode_frame,
511 .long_name = NULL_IF_CONFIG_SMALL("libamr-nb Adaptive Multi-Rate (AMR) Narrow-Band"),
514 AVCodec libamr_nb_encoder =
516 "libamr_nb",
517 CODEC_TYPE_AUDIO,
518 CODEC_ID_AMR_NB,
519 sizeof(AMRContext),
520 amr_nb_encode_init,
521 amr_nb_encode_frame,
522 amr_nb_encode_close,
523 NULL,
524 .sample_fmts = (enum SampleFormat[]){SAMPLE_FMT_S16,SAMPLE_FMT_NONE},
525 .long_name = NULL_IF_CONFIG_SMALL("libamr-nb Adaptive Multi-Rate (AMR) Narrow-Band"),
528 #endif
530 /* -----------AMR wideband ------------*/
531 #if CONFIG_LIBAMR_WB
533 #ifdef _TYPEDEF_H
534 //To avoid duplicate typedefs from typedef in amr-nb
535 #define typedef_h
536 #endif
538 #include <amrwb/enc_if.h>
539 #include <amrwb/dec_if.h>
540 #include <amrwb/if_rom.h>
542 /* Common code for fixed and float version*/
543 typedef struct AMRWB_bitrates
545 int rate;
546 int mode;
547 } AMRWB_bitrates;
549 static int getWBBitrateMode(int bitrate)
551 /* make the correspondance between bitrate and mode */
552 AMRWB_bitrates rates[]={ {6600,0},
553 {8850,1},
554 {12650,2},
555 {14250,3},
556 {15850,4},
557 {18250,5},
558 {19850,6},
559 {23050,7},
560 {23850,8},
562 int i;
564 for(i=0;i<9;i++)
566 if(rates[i].rate==bitrate)
568 return rates[i].mode;
571 /* no bitrate matching, return an error */
572 return -1;
576 typedef struct AMRWBContext {
577 int frameCount;
578 void *state;
579 int mode;
580 Word16 allow_dtx;
581 } AMRWBContext;
583 static int amr_wb_encode_init(AVCodecContext * avctx)
585 AMRWBContext *s = avctx->priv_data;
587 s->frameCount=0;
589 if(avctx->sample_rate!=16000)
591 av_log(avctx, AV_LOG_ERROR, "Only 16000Hz sample rate supported\n");
592 return -1;
595 if(avctx->channels!=1)
597 av_log(avctx, AV_LOG_ERROR, "Only mono supported\n");
598 return -1;
601 if((s->mode=getWBBitrateMode(avctx->bit_rate))<0)
603 av_log(avctx, AV_LOG_ERROR, wb_bitrate_unsupported);
604 return -1;
607 avctx->frame_size=320;
608 avctx->coded_frame= avcodec_alloc_frame();
610 s->state = E_IF_init();
611 s->allow_dtx=0;
613 return 0;
616 static int amr_wb_encode_close(AVCodecContext * avctx)
618 AMRWBContext *s = avctx->priv_data;
620 E_IF_exit(s->state);
621 av_freep(&avctx->coded_frame);
622 s->frameCount++;
623 return 0;
626 static int amr_wb_encode_frame(AVCodecContext *avctx,
627 unsigned char *frame/*out*/, int buf_size, void *data/*in*/)
629 AMRWBContext *s = avctx->priv_data;
630 int size;
632 if((s->mode=getWBBitrateMode(avctx->bit_rate))<0)
634 av_log(avctx, AV_LOG_ERROR, wb_bitrate_unsupported);
635 return -1;
637 size = E_IF_encode(s->state, s->mode, data, frame, s->allow_dtx);
638 return size;
641 static int amr_wb_decode_init(AVCodecContext * avctx)
643 AMRWBContext *s = avctx->priv_data;
645 s->frameCount=0;
646 s->state = D_IF_init();
648 amr_decode_fix_avctx(avctx);
650 if(avctx->channels > 1)
652 av_log(avctx, AV_LOG_ERROR, "amr_wb: multichannel decoding not supported\n");
653 return -1;
656 return 0;
659 static int amr_wb_decode_frame(AVCodecContext * avctx,
660 void *data, int *data_size,
661 AVPacket *avpkt)
663 const uint8_t *buf = avpkt->data;
664 int buf_size = avpkt->size;
665 AMRWBContext *s = avctx->priv_data;
666 const uint8_t*amrData=buf;
667 int mode;
668 int packet_size;
669 static const uint8_t block_size[16] = {18, 23, 33, 37, 41, 47, 51, 59, 61, 6, 6, 0, 0, 0, 1, 1};
671 if(buf_size==0) {
672 /* nothing to do */
673 return 0;
676 mode = (amrData[0] >> 3) & 0x000F;
677 packet_size = block_size[mode];
679 if(packet_size > buf_size) {
680 av_log(avctx, AV_LOG_ERROR, "amr frame too short (%u, should be %u)\n", buf_size, packet_size+1);
681 return -1;
684 s->frameCount++;
685 D_IF_decode( s->state, amrData, data, _good_frame);
686 *data_size=320*2;
687 return packet_size;
690 static int amr_wb_decode_close(AVCodecContext * avctx)
692 AMRWBContext *s = avctx->priv_data;
694 D_IF_exit(s->state);
695 return 0;
698 AVCodec libamr_wb_decoder =
700 "libamr_wb",
701 CODEC_TYPE_AUDIO,
702 CODEC_ID_AMR_WB,
703 sizeof(AMRWBContext),
704 amr_wb_decode_init,
705 NULL,
706 amr_wb_decode_close,
707 amr_wb_decode_frame,
708 .long_name = NULL_IF_CONFIG_SMALL("libamr-wb Adaptive Multi-Rate (AMR) Wide-Band"),
711 AVCodec libamr_wb_encoder =
713 "libamr_wb",
714 CODEC_TYPE_AUDIO,
715 CODEC_ID_AMR_WB,
716 sizeof(AMRWBContext),
717 amr_wb_encode_init,
718 amr_wb_encode_frame,
719 amr_wb_encode_close,
720 NULL,
721 .sample_fmts = (enum SampleFormat[]){SAMPLE_FMT_S16,SAMPLE_FMT_NONE},
722 .long_name = NULL_IF_CONFIG_SMALL("libamr-wb Adaptive Multi-Rate (AMR) Wide-Band"),
725 #endif //CONFIG_LIBAMR_WB